Conviva Connect MCP

 

Conviva Connect MCP (Model Context Protocol) exposes Conviva data to MCP agents so Conviva customers can quickly and securely request Conviva data for MCP agent tasks. Whether providing a Support team with issue context before a Help Desk conversation commences with viewer operational data or automating operational performance reports, Conviva Connect MCP provides the foundation for extremely rich AI applications powered by Conviva data.

MCP is a common standard for Agents to interact with remote data sets. Conviva's MCP implementation builds on existing Conviva APIs, across VSI and DPI, which deliver time series, per-session, per-viewer, real-time and historical aggregated data. Through MCP, AI Agents can discover, query, and consume Conviva datasets dynamically. Agent requests can be built quickly using natural language queries, with insight inferred from the data automatically.

With MCP request-generated dashboards, leverage the power of using Conviva VSI and Conviva DPI data sources compiling application metrics, AI alerts, and streaming sessions based on instructive agent requests, e.g., automated viewer operational analysis based on custom support inquires and operational performance reports for service level (SLO/SLI) compliance.

Viewer Operational Profile Operational Performance Report

Click to expand image

     
     

Click to expand images

Note: This guide provides examples using Claude Desktop, Cursor, and ChatGPT with an MCP Agent running locally within an organization. Conviva also supports using other Large Language Models (LLMs) and LLM clients that support the MCP Interface. Reach out to Conviva Support for details about LLM and LLM client support.

Conviva MCP Servers

Conviva MCP servers provide access to Conviva VSI and Conviva DPI data sources with endpoints for application metrics, AI alerts, and streaming sessions.

Endpoint (Data Source) Application Data

https://mcp.conviva.com/vsi/metrics

VSI

Real-time and historic aggregate streaming metrics.

See Metric Dictionary.

https://mcp.conviva.com/vsi/ai-alerts

VSI

Automated anomaly-based alert noticifications along with the sessions that attributed to the dimension (or set of dimensions) associated with the root cause of the alerts and the impacted viewers.

See AI Alerts.

https://mcp.conviva.com/vsi/sessions VSI

Ssession data for each Conviva streaming session and viewer experience details for the video content based on the viewer ID or viewer IP address.(currently supports only the Viewer Module data)

See Viewer Module.

https://mcp.conviva.com/dpi/metrics

DPI

Real-time and historic application metrics for performance monitoring and user behavior analysis across your web and native applications.

See DPI Metrics.

https://mcp.conviva.com/dpi/ai-alerts

DPI

Automated anomaly-based alert noticifications along with the sessions that attributed to the dimension (or set of dimensions) associated with the root cause of the alerts and the impacted viewers.

See AI Alerts.

 

For the latest release updates and changes, see What's New in Conviva MCP.

LLM Setups

MCP with Claude (Team and Enterprise Level)

Most common enterprise Claude implementations use the Teams or Enterprise levels. For these levels, typically an admin adds the Conviva settings to the claude_desktop_config.json file so other users just install the Claude app and connect to the Conviva MCP servers using their SSO access.

  1. For Team and Enterprise level access, admins enter the Conviva MCP servers data in the claude_desktop_config.json file in the \Claude folder to enable Okta SSO access for all users.

    Copy
    {
      "mcpServers": {
        "conviva-dpi-metrics": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.conviva.com/dpi/metrics"
          ]
        },
        "conviva-vsi-metrics": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.conviva.com/vsi/metrics"
          ]
        },
        "conviva-vsi-sessions": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.conviva.com/vsi/sessions"
          ]
        },
        "conviva-vsi-ai-alerts": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.conviva.com/vsi/ai-alerts"
          ]
        },
        "conviva-dpi-ai-alerts": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.conviva.com/dpi/ai-alerts"
          ]
        }
      }
    }
  2. In the Claude desktop, users update the Settings > Connectors list to connect to the Conviva MCP servers.

    After server authentication prompts, the Claude chat window appears.

MCP with Claude (User Level)

Prerequisites

  • Anthropic Claude Desktop app

  • Node.js

    In the terminal or command prompt, enter npm -v to confirm that Node.js is installed. If Node.js is not installed, 'install it. Node.js is required to run the npm (Node Package Manager) npx command that configures the connectors with the Conviva MCP servers.

    To install Node.js, go to Download Node.js® and copy the install commands into your terminal.

Getting Started

  1. Open the pre-installed Claude desktop app and click Settings at the bottom left corner.

  2. Select Developer and click Edit Config to edit the config file.

    This opens the Claude Settings directory in Finder.

  3. Add the Conviva settings to the claude_desktop_config.json file:

    1. Open the claude_desktop_config.json file using a text editor.

    2. For OAuth authenication to enable Okta SSO access, enter the Conviva MCP servers data.

      Copy
      {
        "mcpServers": {
          "conviva-dpi-metrics": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/dpi/metrics"
            ]
          },
          "conviva-vsi-metrics": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/metrics"
            ]
          },
          "conviva-vsi-sessions": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/sessions"
            ]
          },
          "conviva-vsi-ai-alerts": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/ai-alerts"
            ]
          },
          "conviva-dpi-ai-alerts": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/dpi/ai-alerts"
            ]
          }
        }
      }
    3. For API credential access, replace the AUTH_HEADER string with the Base64-encoded API credential of the desired Conviva c3 account.

      If needed, create API credentials using the Pulse API Management page. It must begin with Basic, as the Conviva API system only supports HTTP basic access authentication.

      Note: To encode API credentials to Base64 format, enter the prompt in the command line:
      echo -n "{client_id}:{client_secret}" | base64
      (replace {client_id} and {client_secret} with the API credentials of the c3 account)

      Copy
      {
        "mcpServers": {
          "conviva-vsi-ai-alerts": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/ai-alerts",
              "--header",
              "Authorization:${AUTH_HEADER}"
            ],
            "env": {
              "AUTH_HEADER": "Basic OVJZaVY3Yn"
            }
          },
          "conviva-vsi-metrics": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/metrics",
              "--header",
              "Authorization:${AUTH_HEADER}"
            ],
            "env": {
              "AUTH_HEADER": "Basic OVJZaVY3Yn"
            }
          },
          "conviva-vsi-sessions": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/vsi/sessions",
              "--header",
              "Authorization:${AUTH_HEADER}"
            ],
            "env": {
              "AUTH_HEADER": "Basic OVJZaVY3Yn"
            }
          },
          "conviva-dpi-metrics": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/dpi/metrics",
              "--header",
              "Authorization:${AUTH_HEADER}"
            ],
            "env": {
              "AUTH_HEADER": "Basic OVJZaVY3Yn"
            }
          },
          "conviva-dpi-ai-alerts": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://mcp.conviva.com/dpi/ai-alerts",
              "--header",
              "Authorization:${AUTH_HEADER}"
            ],
            "env": {
              "AUTH_HEADER": "Basic OVJZaVY3Yn"
            }
          }
        }
      }

      For more details about mcp-remote, see npm: mcp-remote.

  4. In the command line, enter cmd + q to completely quit and restart Claude. In the Windows app, exit and restart the Claude application.

    For OAuth authentication, the browser displays the Conviva authentication page (https://auth.conviva.com/). Enter your email, and your company’s Okta domain login page appears. After a successful login, the browser shows a message Authorization successful message.

    Note: Claude needs a complete restart to apply the configuration changes. To log out of an OAuth session if you are using the NodeJS mcp-remote package, in the command line clear the ~/.mcp-auth directory with rm -rf ~/.mcp-auth. In the Windows app, exit the application. Also, delete any files in the \Users\<user_name>\AppData\Local\npm-cache\_npx folder to refresh the authentication. Then, restart Claude.

  5. Conviva MCPs appear in Claude to enable/disable the desired MCP servers. Each MCP server has tools, such as, Get and List actions for Conviva VSI AI alerts.


  6. Enter a prompt for Conviva data.

    Copy
    get me a summary of my content ai alerts of the last 14 days

     

 

MCP with Cursor

 

Prerequisites

Cursor access

Steps for Cursor Setup

  1. Open a new Cursor window

  2. Click the Settings icon for settings and search for MCP; select the MCP option from the search result list.

  3. Click the Add Custom MCP button to open the mcp.json file, the configuration file for setting up MCPs.


  4. For authorization using API access, copy-paste the following into your mcp.json file, and save the changes.

    Replace {{base64_encoded_api_credential}} with the API credential of the c3 account. For details, see API Management.

    Copy

    {

    {
      "mcpServers": {
        "conviva-ai-alerts": {
          "url": "https://mcp.conviva.com/vsi/ai-alerts",
          "headers": {
            "AUTHORIZATION": "Basic {{base64_encoded_api_credential}}"
          }
        },
        "conviva-vsi-metrics": {
          "url": "https://mcp.conviva.com/vsi/metrics",
          "headers": {
            "AUTHORIZATION": "Basic {{base64_encoded_api_credential}}"
          }
        },
        "conviva-vsi-sessions": {
          "url": "https://mcp.conviva.com/vsi/sessions",
          "headers": {
            "AUTHORIZATION": "Basic {{base64_encoded_api_credential}}"
          }
        },
        "conviva-dpi-metrics": {
          "url": "https://mcp.conviva.com/dpi/metrics",
          "headers": {
            "AUTHORIZATION": "Basic {{base64_encoded_api_credential}}"
          }
        }
      }
    }

    For OAuth access, authenticate using an IDP (Okta) account by clicking on the “Connect“ button.

    Click Open to allow your browser to open the external website for Okta authentication when prompted.

  5. On the Cursor MCP settings page, a green selector indicates that the MCP servers are enabled. Toggle the corresponding switches to select/de-select and enable/disable MCP servers.

  6. Click the “AI Pane” button to open the chat UI and choose the Cursor LLM.

  7. After first-time MCP access, Cursor asks for permission to use the newly added MCP tools. Add the tool to the “Allowlist“ or click Run for each access.

  8. Run data queiies to access Conviva data.

    This example lists the latest AI alerts.

    Copy
    >Give me a summary of the AI alerts for the last 14 days. 

    I'll help you get a summary of your content AI alerts from the last 14 days. Let me fetch this information for you.

    [1 tool called]

    Based on the content AI alerts data from the last 14 days, here's a summary:

    Content AI Alerts Summary (September 9-23, 2025)
    Overall Statistics
    Total Alerts: 543 content AI alerts

    Time Period: September 9, 2025 00:00:00 UTC to September 23, 2025 23:59:59 UTC

    Customer: Demo-Olympics

    Alert Status: All alerts have ended (no active alerts)

    Alert Types by Metric
    Connection Induced Rebuffering Ratio (CIRR): 419 alerts (77.2%)

    Video Start Failures (VSF): 75 alerts (13.8%)  

    Video Playback Failures (VPF): 41 alerts (7.6%)

    Exit Before Video Starts (EBVS): 8 alerts (1.5%)

    Alert Severity Distribution
    Info: 494 alerts (91.0%)

    Warning: 36 alerts (6.6%) 

    Critical: 13 alerts (2.4%)

    Key Patterns Observed
    Daily Peak Activity
    Most alerts occurred during 07:25-07:44 UTC timeframe across multiple days

    This suggests a consistent daily pattern of issues during peak viewing hours

    Root Cause Analysis
    Top Root Causes:

    Content-specific issues: Various Olympic events (Biathlon, Freestyle Skating, Ice Hockey, etc.)

    CDN-related: INHOUSE CDN showing critical severity issues

    Content type: VoD (Video on Demand) content showing multiple critical alerts

    Platform-specific: Roku devices and Olympics-branded apps showing recurring issues

    Critical Alerts
    The 13 critical alerts were primarily related to:

    CIRR (Connection Induced Rebuffering) on INHOUSE CDN

    CIRR for VoD content

    EBVS (Exit Before Video Starts) for live content and AKAMAI CDN

    Geographic and Technical Insights
    Issues span multiple devices (Roku, Apple iPhone, Android)

    Various CDN providers affected (AKAMAI, INHOUSE, EDGECAST)

    Both live and VoD content experiencing issues

    Multiple Olympic channels and events affected

    This data suggests the Demo-Olympics platform experienced consistent but manageable video quality issues during this 14-day period, with most being informational-level alerts that were resolved relatively quickly.

 

 

Prerequisites

ChatGPT access

Steps for ChatGPT Setup

  1. Start ChatGPT using the web interface or desktop app.

  2. With ChatGPT started, at the bottom left corner of the UI click on your username and select Settings.

  3. On the settings pop up, select Apps & connectors.

  4. In the list of Conviva connectors, select the connectors (MCP servers) to activate.

    For example, activating “Conviva VSI Metrics”. Click on the Connect button, then Continue to Conviva VSI Metrics button.



  5. For OAuth access, a series of OAuth steps lead to login. ChatGPT desktop app may pop up a blank page and authenticate without a prompt if a validated Okta session is active.

    For the web UI, a web page refresh may be required to display the newly enabled MCP connector.

  6. After connection, the Conviva MCP server appears in the Enabled apps & connectors list.

  7. Start a new chat session and prompt ChatGPT for a list of c3 accounts.

    ChatGPT may prompt for a confirmation to use the MCP tools. Choose Allow for conversation and confirm.

  8. Query your Conviva data using ChatGPT.

  9. If you have access to multiple .c3 accounts, also prompt ChatGPT to compare data across accounts.

 

 

Conviva Connect MCP Claude Cursor ChatGPT Claude Cursor ChatGPT Conviva Connect MCP Claude Conviva MCP